In a cluster system in which an active system device and a standby system device are connected by a network, at the time of a failure, fault, or periodic maintenance of the active system device, service processing performed on the active system device is taken over by the standby system device. However, because there is no guarantee that the takeover is carried out without fail, the reliability of the system decreases.
There are some factors which cause the failure of the takeover of the service processing. The biggest factor in this is a shortage of resources in the standby system device. In order to take over the service processing performed in the active system device by the standby system device, the standby system device has to acquire an amount of resources that is greater than the amount of resources (for example, a memory amount, a disk capacity, or the like) that is used in the active system device to perform the service processing. However, in the cluster system that is not configured as a hot standby, at a time at which the takeover is required, a process for the takeover is activated in the standby system device and the resources are acquired. Therefore, a status in which a sufficient amount of resources cannot be acquired occurs because of the shortage of resources in the standby system device at the time.
Accordingly, in order to prevent a failure of the takeover of the service processing caused by the shortage of resources in the standby system device, the following technology is proposed (for example, refer to Japanese Patent Publication No. 3248485). First, the active system device periodically detects usage of resources and notifies the standby system device. The standby system device compares the usage of the resources that is reported with an available amount of resources of the standby system device. When the available amount of resources of the standby system device is smaller than the usage of the resources of the active system device that is reported, the standby system device determines that the fault occurs and transmits a report showing abnormality.